Wokelo MCP Server

Dealmaking research for AI assistants - company intelligence, transaction data, and research deliverables via MCP.

Website · Documentation

The problem with AI assistants for dealmaking research

You are likely already using an AI assistant - Claude, ChatGPT, Copilot, or similar. But every one of them shares the same core limitations: training data with a cutoff date, no access to premium sources, a reliance on web search that surfaces the same SEO-optimized headlines everyone else sees, and the need for constant back-and-forth prompting just to get to a usable result.

How Wokelo MCP supercharges your existing tools

Wokelo MCP is built to solve every one of those limitations for dealmaking research - bundled premium subscriptions, a native company and transactions database, zero hallucination, a proprietary news algorithm to surface signals your competitors aren't seeing, and a single query that delivers a finished deliverable without you staying in the loop.

Quickstart

Wokelo MCP is a hosted remote server. No local installation required. Connect using the server URL and authenticate once via browser.

Server URL:

https://mcp.wokelo.ai/mcp

Transport: Streamable HTTP Authentication: OAuth 2.0 with Dynamic Client Registration (browser-based login on first connection)

Claude Desktop (Recommended)

Claude Desktop connects to remote OAuth servers through its built-in Connectors UI:

1. Open Claude Desktop
2. Go to Settings > Connectors
3. Click Add MCP Server
4. Paste the server URL: https://mcp.wokelo.ai/mcp
5. Complete the OAuth login in your browser
6. All 25 Wokelo tools will appear automatically

Note: Remote MCP servers with OAuth authentication are added through the Connectors UI, not through the claude_desktop_config.json file. This is by design in Claude Desktop.

Claude Code

Run the following in your terminal:

claude mcp add --transport http wokelo https://mcp.wokelo.ai/mcp

Then open a Claude Code session and run /mcp to complete the OAuth authentication flow. Your browser will open for sign-in.

Verify: Run /mcp again — wokelo should appear as connected with its tool list.

ChatGPT Web

Availability: Full MCP support is available on ChatGPT web for Business and Enterprise/Edu workspaces. Only admins/owners can publish apps, and developer mode must be enabled before creating a custom MCP app.

1. Confirm developer mode is enabled for your account:
   • For Business, admins/owners can enable it when creating a custom app from Workspace Settings > Apps > Create.
   • For Enterprise/Edu, admins can grant access first, then enabled users can turn it on in Settings > Apps > Advanced Settings.
2. Open Workspace Settings > Apps > Create.
3. Add the app details:
   • Name: Wokelo
   • MCP Server URL: https://mcp.wokelo.ai/mcp
   • Authentication: OAuth
4. Click Create.
5. Complete the OAuth login in your browser when redirected to Wokelo.
6. After authentication, the Wokelo app appears in ChatGPT as a draft/dev app for testing.

Optional: If you want the connector available to the full workspace, publish it from Workspace Settings > Apps after testing.

OpenAI Codex (CLI and IDE extension)

Codex shares MCP configuration between the CLI and the IDE extension via ~/.codex/config.toml.

Step 1 — Add the server (either command or direct file edit):

codex mcp add wokelo --url https://mcp.wokelo.ai/mcp

Or add to ~/.codex/config.toml directly:

[mcp_servers.wokelo]
url = "https://mcp.wokelo.ai/mcp"

Step 2 — Complete OAuth sign-in:

codex mcp login wokelo

This opens your browser for authentication. Codex stores the tokens locally and refreshes them automatically. The codex mcp login command works only for Streamable HTTP servers that advertise OAuth — which Wokelo does.

Verify: Start a Codex session (codex) and run /mcp in the TUI. Confirm wokelo lists its tools. The IDE extension reflects the same configuration automatically.

Cursor

Cursor 1.0+ supports OAuth and Streamable HTTP natively:

1. Go to Cursor > Settings > Cursor Settings > MCP
2. Click Add MCP Server
3. Paste the server URL: https://mcp.wokelo.ai/mcp
4. Complete the OAuth login when prompted
5. Click the refresh icon in the MCP panel if the tool list doesn't appear immediately

Microsoft Copilot Studio

Copilot Studio has a built-in MCP onboarding wizard that handles OAuth 2.0 with Dynamic Client Registration (DCR) — which Wokelo supports — so no Swagger file or custom connector is required.

1. Open your agent in Copilot Studio → go to the Tools page
2. Select Add a tool → New tool → Model Context Protocol
3. In the onboarding wizard, fill in:
   • Server name: Wokelo
   • Server description: Dealmaking research — company intelligence, transactions, and research deliverables
   • Server URL: https://mcp.wokelo.ai/mcp
4. Under Authentication, select OAuth 2.0
5. For the OAuth type, select Dynamic discovery
6. Click Create, then Next
7. On the Add tool dialog, select Create a new connection → complete the browser sign-in
8. Click Add to agent

The agent orchestrator uses the server description to decide when to call Wokelo tools, so keep the description clear and specific.

Note: Generative Orchestration must be enabled on the agent for MCP tools to be invoked. Copilot Studio dropped SSE support in August 2025; Streamable HTTP (what Wokelo uses) is the supported transport.

VS Code / GitHub Copilot

VS Code supports remote MCP servers over Streamable HTTP and can add them through the built-in guided flow:

1. Open the Command Palette and run MCP: Add Server
2. Choose whether to add the server to your Workspace or your User profile
3. Enter the server name: wokelo
4. Select HTTP as the server type
5. Enter the server URL: https://mcp.wokelo.ai/mcp
6. Save the configuration and start the server when prompted
7. If VS Code asks you to trust the server configuration, confirm trust
8. Complete the OAuth sign-in in your browser when VS Code prompts for authentication
9. Open Copilot Chat, switch to Agent mode, and use the tools picker / Configure Tools to confirm the Wokelo tools are available

If you prefer to manage the config directly, add this to .vscode/mcp.json or your user mcp.json:

{
  "servers": {
    "wokelo": {
      "type": "http",
      "url": "https://mcp.wokelo.ai/mcp"
    }
  }
}

After saving the file, start the server from the editor CodeLens or via MCP: List Servers.

GitHub Copilot CLI

1. Install GitHub Copilot CLI globally. With npm:

npm install -g @github/copilot

2. Start the CLI:

copilot

3. Sign in to GitHub by running:

/login

You can also authenticate from the terminal with copilot login.

4. In the interactive Copilot CLI session, run:

/mcp add

5. Fill in the form with:
   • Server Name: wokelo
   • Server Type: HTTP or SSE
   • URL: https://mcp.wokelo.ai/mcp
   • Tools: *
6. Press Ctrl+S to save the configuration
7. Complete the Wokelo OAuth login in your browser when prompted

Verify: Run /mcp show wokelo

Perplexity (Pro / Enterprise)

Custom remote connectors are available to paid Perplexity users and support OAuth 2.0 with Dynamic Client Registration natively.

1. Open Perplexity → Settings → Connectors
2. Click + Custom connector in the top-right, then select Remote
3. Fill in:
   • Name: Wokelo
   • MCP Server URL: https://mcp.wokelo.ai/mcp
   • Transport: Streamable HTTP
   • Authentication: OAuth 2.0 — leave Client ID and Client Secret blank (Wokelo supports DCR, so endpoints are auto-discovered via /.well-known/oauth-protected-resource/mcp)
4. Check the acknowledgement box and save
5. Complete the browser sign-in when prompted
6. Toggle the Wokelo connector on under Sources on the Perplexity home screen

Enterprise note: Organization-wide sharing requires an admin to share the connector from the Permissions screen after creation. With OAuth, admins can either authenticate once on behalf of the organization or require each member to authenticate individually.

Windsurf

1. Open Windsurf Settings > MCP
2. Add a new server with URL: https://mcp.wokelo.ai/mcp
3. Complete the OAuth login when prompted

Visual Studio 2022 (17.14+) / Visual Studio 2026

Visual Studio supports remote MCP servers with OAuth via an mcp.json file at the solution or user level:

{
  "servers": {
    "wokelo": {
      "type": "http",
      "url": "https://mcp.wokelo.ai/mcp"
    }
  }
}

Use the Manage Authentication CodeLens action in the mcp.json file to trigger the OAuth browser flow.

Google Antigravity

Note: Antigravity's native handling of OAuth with remote MCP servers has been inconsistent in publicly reported cases. Try the native config first; if the OAuth flow fails or hangs, use the mcp-remote bridge fallback.

Open Agent pane → … menu → MCP Servers → Manage MCP Servers → View raw config.

Option A — Native config (try first):

{
  "mcpServers": {
    "wokelo": {
      "serverUrl": "https://mcp.wokelo.ai/mcp"
    }
  }
}

Save, click Refresh, and complete OAuth when prompted.

Option B — mcp-remote bridge (fallback, requires Node.js):

{
  "mcpServers": {
    "wokelo": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote@latest",
        "https://mcp.wokelo.ai/mcp"
      ]
    }
  }
}

mcp-remote handles the OAuth flow in your system browser and proxies the connection to Antigravity as a local stdio server.

Any Other MCP Client

Wokelo works with any MCP client that supports Streamable HTTP transport and OAuth 2.0. Point your client at:

https://mcp.wokelo.ai/mcp

Full connection details:

On first connection, your client opens a browser tab for authentication. After sign-in, tokens are stored by your client and refreshed automatically.

If your client only supports stdio (not Streamable HTTP): use mcp-remote as a bridge:

{
  "mcpServers": {
    "wokelo": {
      "command": "npx",
      "args": ["-y", "mcp-remote@latest", "https://mcp.wokelo.ai/mcp"]
    }
  }
}

Verifying the server before connecting a client

To confirm the server is healthy independently of any client, use the official MCP Inspector:

npx @modelcontextprotocol/inspector

In the Inspector UI, choose Streamable HTTP, enter https://mcp.wokelo.ai/mcp, click Connect, and sign in. If Inspector lists tools under the Tools tab, the server is working correctly — any compliant client should also work.

Tools

Currently, Wokelo MCP exposes 25 tools across two categories.

Read-Only Tools (16)

Write Tools (9)

Data Sources

Wokelo bundles 30+ premium data sources so you don't manage separate subscriptions:

Use Cases

Company Enrichment - Core Datasets

"For Cleo (meetcleo.com) - do a data-backed analysis across firmographics, funding rounds and investors, recent news, UK financials, employee trends, products, website traffic."

Your AI assistant calls Get Single Company Enrichment, pulls verified data across all relevant dimensions, and returns a structured briefing in one turn. Works for well-known and obscure companies alike - structured data is aggregated across sources that don't talk to each other (e.g. one funding round from PitchBook and one from a news article).

Company Enrichment - Bulk

"Find the top LLM observability startups in the US who raised funding in the last 3 years. Enrich each with funding stage, amount raised, key investors, and key products. Output an Excel file."

Uses Search Company to identify the universe, then Submit Company Enrichment to enrich every row in a single pass. No re-prompting per row. Results compound in bulk - LLMs struggle when dealing with multiple companies and multiple structured datasets simultaneously.

Company Enrichment - Earnings and Filings

"Look at Snowflake's last 4 earnings calls. Extract these themes from each call in a grid view: AI commentary, product launches, management changes, financial outlook, risks. Add exact quote excerpts for each cell."

Uses Get Earnings Calls and List Earnings Calls to pull structured transcripts from proprietary datasets (including S&P). Raw transcripts enable nuanced analysis for granular insights that get missed in open-ended Q&A without Wokelo.

Company Enrichment - Alternative Datasets

"Look at Glassdoor reviews of Ensemble Health Partners. Prepare an analysis of gaps and improvement areas. Also prepare an interactive UI to view review summary and all individual reviews."

Uses Get Employee Reviews and Get Product Reviews to pull all raw review data. Getting all raw data is typically difficult for an LLM, and without raw data, granular insights are patchy. Especially useful for portfolio teams and product teams tracking individual reviews.

Company Discovery - Market Map

"I am looking at flooring companies in US - founded at least 10 years ago, with focus on manufacturers. I need a long list of companies with details and analytics of players."

Submit Market Map generates comprehensive market maps in seconds. LLMs typically produce only 30-40 companies max without hallucinations. Wokelo's market map also enriches each company with structured data and supports filters for geography, employees, type, funding, and more.

M&A Target Screening

"We're looking at acquisition targets in construction tech vertical SaaS. Find companies with $5-50M revenue, Series A-C, strong product reviews."

Submit Target Screening identifies targets matching your criteria. Follow up with Submit Company Intelligence for deep dives on the shortlist.

Buyer Identification

"Who are the most likely acquirers for a mid-market cybersecurity MSSP?"

Submit Buyer Screening identifies potential acquirers based on strategic fit, deal history, and sector alignment.

Hybrid - Buyer Screening + Portfolio Analysis

"For Warburg Pincus - prepare a dashboard of all M&A and investments. Create themes around portcos nearing their term. Select a sample portco nearing harvest period and find 30 top buyers with key metrics and rationale."

This is an example of a hybrid query made possible by Wokelo's width of data and insights. Leverages curated algorithms like buyer screening + target screening alongside structured M&A and investment data not available via web search.

News Analysis - On-Demand

"Find all news related to lawsuits and litigations related to Gen-AI. Look at news of OpenAI, Anthropic, Google, Microsoft, Cohere, Perplexity. Summarize in an insightful table."

Get Company News provides real-time, deduplicated, time-series news with categorization and date filtering. Can be used for nuanced analysis of specific news types across multiple companies for cluster, peer set, or industry analysis. LLM web search is generally not exhaustive for periodic news.

News Analysis - Scheduled Monitoring

"Every Monday morning, send me a digest of key developments across our portfolio companies from the prior week."

Set up as a recurring scheduled task (via Claude Cowork or similar). Build your own news monitoring dashboard that runs on a cadence of your choosing - daily, hourly, weekly. Get Company News monitors thousands of publishers and returns deduplicated, entity-resolved intelligence on each run.

Peer Sentiment Analysis

"We have a value creation review for Make.com next week. Pull employee and software review data for them and their three closest peers. Compare culture scores and leadership ratings. Flag themes."

Uses Submit Company Peers to identify peers, then Get Employee Reviews and Get Product Reviews to pull alt-data from Glassdoor and G2 that a generic LLM cannot reliably access.

Trigger Wokelo Reports

"Do a peer comparison workflow comparing OpenAI and Anthropic in Wokelo."

Start Workflow triggers pre-built workflows without needing to visit the Wokelo platform. Can trigger frequently used workflows or custom workflows your org has access to. After completion, use the structured JSON output to create deliverables in your format of choice (3-page Word doc, 5-page PPT, etc.).

Examples

See Wokelo MCP in action with real outputs:

Browse all sample outputs

How It Works

┌─────────────────┐     MCP (Streamable HTTP)     ┌──────────────────┐
│                  │ <──────────────────────────── │                  │
│  Your AI Client  │         OAuth 2.0             │  Wokelo Platform │
│  (Claude, etc.)  │ ────────────────────────────> │                  │
│                  │                                │  100M+ companies │
└─────────────────┘                                │  30+ data sources│
                                                   │  Transaction DB  │
  You ask a question                               │  News algorithm  │
  |                                                │  Workflow engine │
  Your AI calls Wokelo tools                       └──────────────────┘
  |                                                        |
  Wokelo returns verified,                                 |
  structured, source-cited data                  ──────────┘
  |
  Your AI synthesizes the answer

1. Connect - Add the server URL to your MCP client. Authenticate once via OAuth.
2. Ask - Query your AI assistant in natural language. It determines which Wokelo tools to call.
3. Receive - Wokelo returns verified, structured data. Your assistant synthesizes and presents it.
4. Async support - Long-running jobs (enrichment, workflows) run in the background. Use Poll Request to Completion or Wait for Request Result to pick up results when ready.

Requirements

• An MCP client that supports Streamable HTTP transport and OAuth 2.0
• A Wokelo account - sign up at wokelo.ai
• No local installation, packages, or API keys to manage

Supported Clients

Troubleshooting

To isolate whether an issue is client-side or server-side, test the server with MCP Inspector (see the verification section above). If Inspector connects and lists tools, the server is fine and the issue is with the client configuration.

Security & Compliance

• OAuth 2.0 - No API keys stored in config files. Browser-based authentication with token refresh.
• SOC 2 Type II Compliant - Independently audited security controls for data handling, availability, and confidentiality.
• ISO 27001 Certified - Information security management system certified to international standards.
• No data leaves Wokelo - Queries are processed on Wokelo's infrastructure. Responses are returned to your MCP client.

Documentation

Full setup guides, API reference, and tutorials:

docs.wokelo.ai/getting-started-mcp

Registry Publishing

This repository includes official MCP Registry metadata in server.json.

• The server is defined as a public remote MCP server at https://mcp.wokelo.ai/mcp
• Metadata is validated automatically through GitHub Actions in .github/workflows/validate-mcp.yml
• Publishing is automated through GitHub Actions in .github/workflows/publish-mcp.yml
• To publish a new MCP Registry version, create and push a Git tag in the format vX.Y.Z

Example:

git tag v1.0.0
git push origin v1.0.0

The workflow uses GitHub OIDC authentication and stamps the registry version from the Git tag before publishing.

Support

• Documentation: docs.wokelo.ai/getting-started-mcp
• Email: sales@wokelo.ai
• Website: wokelo.ai

License

Proprietary. Usage is subject to the Wokelo Terms of Service.

Wokelo AI - Intelligence infrastructure for dealmakers.